ACTUATOR DYNAMICS AND DELAY COMPENSATION USING NEUROCONTROLLERS

This study compares the performances of a conventional feedforward con troller and a neurocontroller in compensating the effects of the actua tor dynamics and computational phase delay in a simple digital vibrati on control system. The model of this system is based on an earlier exp erimental study where a two-degree-of-freedom dynamic system was const ructed in the laboratory. This system consisted of two electrohydrauli c, position-controlled actuator-mass systems, mounted one on top of th e other. Bode frequency domain plot were used to identify the governin g parameters of the system. Based on the identified model of the exper imental setup, a conventional feedforward controller and a neurocontro ller are designed to compensate for the adverse effects of actuator dy namics and computational phase delay. The identified model is used to generate training information for the neurocontroller in a frequency d omain of interest. To accomplish this endeavor, a neural network simul ator is developed, This software uses a modified generalized delta rul e with an adaptive momentum term for its learning mechanism and has a dynamic network topology capability. Through experiments and numerical simulations, it is shown that the neurocontroller is far more effecti ve in compensating for actuator dynamics and time delay than the conve ntional feedforward controller. Factors contributing to the superior p erformance of the neurocontroller are identified and discussed.